The objective of this proposal is to develop and evaluate improved statistical methods for the analysis of complex prospective failure time studies. Major areas of application include epidemiological and clinical cohort studies that aim to relate risk factor levels or risk factor changes over time to subsequent disease incidence, disease mortality or to some other 'failure' time endpoint. Other areas of application include therapeutic trials and certain basic research studies.